1 00:00:00,240 --> 00:00:03,080 Speaker 1: Food Stuff six months facial recognition trial seems to have 2 00:00:03,080 --> 00:00:05,640 Speaker 1: got the tick. The early indications suggests one hundred and 3 00:00:05,680 --> 00:00:08,240 Speaker 1: thirty events like assault and verbal abuse have been avoided. 4 00:00:08,800 --> 00:00:11,119 Speaker 1: They are standing by for a final report. But Foodstuff's 5 00:00:11,119 --> 00:00:15,440 Speaker 1: North Island General Counsel Julian Benfield, Well, it's Julian morning 6 00:00:15,440 --> 00:00:15,760 Speaker 1: to you. 7 00:00:16,480 --> 00:00:17,000 Speaker 2: Morning Mike. 8 00:00:17,320 --> 00:00:18,639 Speaker 1: What have you got I mean when we say the 9 00:00:18,640 --> 00:00:20,520 Speaker 1: final report, what have you got in front of you? 10 00:00:20,880 --> 00:00:22,720 Speaker 1: And how much more was coming in a final report 11 00:00:22,720 --> 00:00:24,080 Speaker 1: that may or may not change anything. 12 00:00:24,960 --> 00:00:27,160 Speaker 2: Yeah, as you say, our trial has come to an end, 13 00:00:27,200 --> 00:00:30,160 Speaker 2: and so this trial of facial recognition technology in twenty 14 00:00:30,160 --> 00:00:33,320 Speaker 2: five stores was all about helping to keep our people safe. 15 00:00:34,040 --> 00:00:36,720 Speaker 2: We've got a draft report from our independent evaluator, so 16 00:00:36,760 --> 00:00:39,040 Speaker 2: that's an evaluator that we appointed based on the advice 17 00:00:39,080 --> 00:00:42,680 Speaker 2: of the Privacy Commission. The early results from them are strong. 18 00:00:43,120 --> 00:00:46,199 Speaker 2: It's showing seventeen hundred and forty seven actual matches of 19 00:00:46,280 --> 00:00:48,680 Speaker 2: repeat offenders. And as you have said that the early 20 00:00:48,720 --> 00:00:52,000 Speaker 2: results show that we've avoided serious incidents, one hundred and 21 00:00:52,040 --> 00:00:54,720 Speaker 2: thirty serious incidents, so that's things like assault, converible abuse. 22 00:00:55,360 --> 00:00:57,200 Speaker 2: The safety of our team is a top priority. So 23 00:00:57,240 --> 00:01:00,000 Speaker 2: avoiding around one hundred and thirty serious incidents is huge. 24 00:01:00,400 --> 00:01:02,160 Speaker 2: That's one hundred and thirty times that our team mem 25 00:01:02,200 --> 00:01:05,280 Speaker 2: design in harm's way or attacked or abused. And we've 26 00:01:05,280 --> 00:01:07,560 Speaker 2: also said that our customers are very supportive of what 27 00:01:07,600 --> 00:01:10,759 Speaker 2: we're doing. We are based on that draft report. We're 28 00:01:10,800 --> 00:01:14,360 Speaker 2: continuing to use facial recognition technology in the meantime while 29 00:01:14,400 --> 00:01:17,000 Speaker 2: we wait for that final report from an independent evaluator. 30 00:01:17,280 --> 00:01:19,839 Speaker 2: But they have said to us that they don't expect 31 00:01:19,880 --> 00:01:22,839 Speaker 2: the findings to materially change between draft report and final report. 32 00:01:22,959 --> 00:01:25,040 Speaker 1: So how do you know it was avoided if it 33 00:01:25,080 --> 00:01:27,080 Speaker 1: didn't happen. How literally does that unfold? 34 00:01:28,120 --> 00:01:31,560 Speaker 2: So we've got our independent evaluator. That factor has come 35 00:01:31,560 --> 00:01:33,160 Speaker 2: from the ends of the one hundred and thirty incidents 36 00:01:33,160 --> 00:01:35,920 Speaker 2: that are avoided has come from independent evaluator Scalety, who's 37 00:01:35,920 --> 00:01:39,480 Speaker 2: a respected research and evaluation firm. What they've looked at 38 00:01:39,680 --> 00:01:43,040 Speaker 2: is harmful behavior data at the twenty five trial stores 39 00:01:43,080 --> 00:01:45,600 Speaker 2: and compared that against the non trial stores. They've looked 40 00:01:45,600 --> 00:01:48,160 Speaker 2: at the direct deffect of facial recognition, so the evidence 41 00:01:48,160 --> 00:01:51,560 Speaker 2: that it's actually detected offenders in store, as well as 42 00:01:51,600 --> 00:01:54,240 Speaker 2: the deterrent effect which has been real, So offenders choosing 43 00:01:54,280 --> 00:01:56,160 Speaker 2: not to come to the store at all because they 44 00:01:56,160 --> 00:01:58,520 Speaker 2: think that the technology will detect them. 45 00:01:58,760 --> 00:02:02,600 Speaker 1: Good bang for buck scenario, the technology to put it in, 46 00:02:02,840 --> 00:02:05,559 Speaker 1: how many you may or may not avoid, et cetera, 47 00:02:05,720 --> 00:02:07,680 Speaker 1: versus the cost of you know what, does it work? 48 00:02:07,720 --> 00:02:10,120 Speaker 1: Does it add up well? 49 00:02:09,680 --> 00:02:13,440 Speaker 2: And these early results absolutely show that it's been a success. 50 00:02:13,520 --> 00:02:16,359 Speaker 2: You know, we've seen some serious and sort of incidents avoided, 51 00:02:16,440 --> 00:02:20,000 Speaker 2: minimal privacy impact based on early findings and customer support. 52 00:02:20,280 --> 00:02:21,960 Speaker 2: In terms of the cost, we don't have a fixed 53 00:02:22,000 --> 00:02:24,040 Speaker 2: budget to tack or retail crime. We can't put a 54 00:02:24,080 --> 00:02:26,519 Speaker 2: price on the safety of our team and customers. So 55 00:02:26,520 --> 00:02:28,280 Speaker 2: it alwyos been what we need to do to address 56 00:02:28,320 --> 00:02:30,079 Speaker 2: such an important issue? Is it? 57 00:02:30,120 --> 00:02:31,920 Speaker 1: Could it be? Do you do I mean you might 58 00:02:32,000 --> 00:02:33,680 Speaker 1: be in the wrong department. Do you have a vibe 59 00:02:33,720 --> 00:02:35,480 Speaker 1: that you know? We were talking to Andrew cost At 60 00:02:35,480 --> 00:02:37,920 Speaker 1: the Police Commissioner before. This is all post COVID and 61 00:02:37,919 --> 00:02:40,040 Speaker 1: we're all angsty and angry and furious and cost of 62 00:02:40,080 --> 00:02:41,840 Speaker 1: living and all that sort of stuff. Is it possible 63 00:02:41,880 --> 00:02:44,600 Speaker 1: this thing will solve itself to a degree. 64 00:02:45,520 --> 00:02:47,360 Speaker 2: Well, we think we need to do more, and that's 65 00:02:47,400 --> 00:02:49,960 Speaker 2: why we've been raising attention to the retail crime issue 66 00:02:50,000 --> 00:02:53,280 Speaker 2: for two US so years now. We've seen massive increases 67 00:02:53,280 --> 00:02:56,760 Speaker 2: in retail crime during that period. It is fair to 68 00:02:56,760 --> 00:02:59,280 Speaker 2: say that there has been an increase in aggression in 69 00:02:59,280 --> 00:03:02,400 Speaker 2: our stores since COVID area. But our experts tell us 70 00:03:02,440 --> 00:03:05,160 Speaker 2: that it's not cost of living that is causing this issue. 71 00:03:05,800 --> 00:03:09,560 Speaker 2: Our security experts tell us it's professional criminals targeting brand 72 00:03:09,600 --> 00:03:12,760 Speaker 2: specific goods and links organized crime and not linked to 73 00:03:12,800 --> 00:03:15,040 Speaker 2: cost of living. And to be honest, no violence or 74 00:03:15,040 --> 00:03:17,440 Speaker 2: aggression and our stores is acceptable. You know, all of 75 00:03:17,440 --> 00:03:20,480 Speaker 2: our store teams deserve to return home safe every day. 76 00:03:20,680 --> 00:03:23,839 Speaker 1: Those social media videos that I used to watch where 77 00:03:23,840 --> 00:03:27,880 Speaker 1: the guy would literally walk through barefoot, normally grab a 78 00:03:28,080 --> 00:03:30,560 Speaker 1: can of beers and a side of beef and then 79 00:03:30,760 --> 00:03:34,200 Speaker 1: just walk out as the security guard watched them do that. 80 00:03:34,320 --> 00:03:37,280 Speaker 1: Does that still happen? And if it does regularly. 81 00:03:38,280 --> 00:03:41,560 Speaker 2: I mean it happens all the time. We've reported on 82 00:03:41,560 --> 00:03:44,480 Speaker 2: our incidents, and while they're slightly down this quarter, there 83 00:03:44,520 --> 00:03:47,480 Speaker 2: are a huge number of incidents that come through. And 84 00:03:47,520 --> 00:03:50,400 Speaker 2: you know, our trial is focused on repeat offenders, and 85 00:03:50,440 --> 00:03:52,600 Speaker 2: in the last quarter, thirty eight percent of all offenses 86 00:03:52,640 --> 00:03:55,800 Speaker 2: were caused by repeat offenders. So this facial recognition technology 87 00:03:55,800 --> 00:03:58,280 Speaker 2: helps us to identify the defenders before they come into 88 00:03:58,320 --> 00:04:01,280 Speaker 2: the store. And it's one of the most proactive technologies 89 00:04:01,320 --> 00:04:04,480 Speaker 2: that we can find because if you catch the offender 90 00:04:04,560 --> 00:04:06,320 Speaker 2: earlier on in the store journey, they tend to be 91 00:04:06,400 --> 00:04:07,880 Speaker 2: less aggressive when they're approached. 92 00:04:08,000 --> 00:04:12,040 Speaker 1: Good conspiracy text, Mike, how many false positives? And how 93 00:04:12,040 --> 00:04:13,800 Speaker 1: do we know they aren't using the data of an 94 00:04:13,800 --> 00:04:16,120 Speaker 1: additional surveillance and shopper tracking. 95 00:04:17,120 --> 00:04:20,360 Speaker 2: Yeah, well on the second point, absolutely not. This We've 96 00:04:20,400 --> 00:04:23,240 Speaker 2: been very clear and transparent to New Zealand about the 97 00:04:23,279 --> 00:04:25,599 Speaker 2: fact that this is solely about retail crime and it 98 00:04:25,680 --> 00:04:28,839 Speaker 2: is solely about keep targeting repeat affendas and stopping them 99 00:04:28,880 --> 00:04:31,560 Speaker 2: from doing more harm. That is all it's about. It's 100 00:04:31,560 --> 00:04:34,320 Speaker 2: not about anything else. That is the sole focus of 101 00:04:34,400 --> 00:04:34,880 Speaker 2: our trial. 102 00:04:35,160 --> 00:04:37,520 Speaker 1: All right, Julian, appreciate it very much. Julian Benefield, who 103 00:04:37,560 --> 00:04:40,599 Speaker 1: is with Foodstuffs in North Island their general counsel. For 104 00:04:40,720 --> 00:04:43,760 Speaker 1: more from the Mic Asking Breakfast, listen live to news 105 00:04:43,800 --> 00:04:46,719 Speaker 1: talks that'd be from six am weekdays, or follow the 106 00:04:46,760 --> 00:04:48,200 Speaker 1: podcast on iHeartRadio.